Prototype testing presents a realistic but non-functional model of a product or feature to users and observes their interactions. It happens before any production code is written, making it cheap to learn and cheap to change.
After ideation and before development, to validate that a proposed design actually works for real users.
- Build a prototype at appropriate fidelity: paper, wireframe, or high-fidelity Figma
- Match fidelity to what you need to learn — interaction patterns need low-fi; visual design needs high-fi
- Define the specific tasks you want to test
- Recruit 5 users from your genuine target segment
- Observe without intervening — note hesitations, errors, and verbal reactions
- Synthesise findings into prioritised design changes before handing off to engineering
High-fidelity Figma prototype of a redesigned 'Your Library'. Task: 'Find the album you added 3 months ago.' Finding: all 5 users immediately look for a search box within Library — but the prototype doesn't have one. They expect search within Library as a first-class feature. This changes the engineering scope before a single line of code is written.
Please contact the author for more information on these examples at linkedin.com/in/kshitijrege
- Testing with people who already know the design — their familiarity invalidates results
- Making the prototype so polished it looks finished — users hold back honest criticism
- Not having a specific task — 'just explore it' produces weak, unactionable insights
- Don't Make Me Think — Steve Krug
- Sprint — Jake Knapp et al.